Study On Layout Of Detector For Urban Dynamic Traffic Information Collection System And Short-Term Traffic Flow Forecasting

Based on the researches home and abroad, the real-time and accuracy of dynamic traffic information is the key of Intelligent Transportation Systems (ITS). However, during the practical application, the development of urban dynamic traffic information collection system is lack of coordinative and theoretical guidance. And the optimal location of traffic detectors is also short of theoretical foundation. Sometimes, the deletion of traffic data occurs due to the limitation of detectors. In this instance, short-term traffic flow forecasting which according to the historical data becomes to be a hotspot.First, the status quo of urban dynamic traffic information collection system home and abroad is stated by a mass of basic study and preparation, and then its frame is designed. Second, intrusive detectors and non-intrusive detectors are introduced respectively. In succession, the study compares the differences between principles of operation, applications, performance, costs and etc. After the analysis above, the commonly principle of the location of traffic detectors are displayed. Where after, the methods and steps of traffic detectors'optimal location are put forward, and applied fuzzy mathematics comprehensive evaluation to traffic detectors'selection.At last, considering the deletion of traffic data, short-term traffic flow forecasting is putting forward by historical data. Through the analysis of varied predictive models, neural networks has many virtues like identifying complicated non-linear systems and self adaptation, and it is one of the most potential models. The RBF neural networks and matlab are chosen for short-term traffic flow forecasting, and the principium and arithmetic of RBF are analyzed with an example.